Enabling-more-agents-for-Batch-RL-and-cleanup

IntelLabs / coach

Reinforcement Learning Coach by Intel AI Lab enables easy experimentation with state of the art Reinforcement Learning algorithms

https://intellabs.github.io/coach/

Apache License 2.0

2.32k stars 459 forks source link

Enabling-more-agents-for-Batch-RL-and-cleanup #258

Closed gal-leibovich closed 5 years ago

gal-leibovich commented 5 years ago

allowing for the last training batch drawn to be smaller than batch_size + adding support for more agents in BatchRL by adding softmax with temperature to the corresponding heads + adding a CartPole_QR_DQN preset with a golden test + cleanups

The list of supported agents in BatchRL is now extended to - DQN DDQN Dueling DDQN Categorical DQN (C51) Rainbow DQN QR DQN Bootstrapped DQN NEC MMC PAL